overdriving / underdriving COB led data?

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
dear all,
one of my favorite LED to use in random projects are the generic 10w Chinese COB. they're under rs 50 a piece and very powerful for what they are.

what puzzles me is absence of any research regarding over /under driving them ?.I have found at length research about the subject on 5mm and 1w- 5w single emmiters. but nothing about COB (Chinese or otherwise ).

anyone has any lead on them?
the ones I use are rated 900ma at about 9-11v. they have 9 dies in a matrix and are easily available on eBay.

thanks.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
dear all,
one of my favorite LED to use in random projects are the generic 10w Chinese COB. they're under rs 50 a piece and very powerful for what they are.

what puzzles me is absence of any research regarding over /under driving them ?.I have found at length research about the subject on 5mm and 1w- 5w single emmiters. but nothing about COB (Chinese or otherwise ).

anyone has any lead on them?
the ones I use are rated 900ma at about 9-11v. they have 9 dies in a matrix and are easily available on eBay.

thanks.

That's because that data should be readily evident in the data sheet .... minus the data on color shift.
 

FRITZHID

Flashlight Enthusiast
Joined
Jun 20, 2011
Messages
2,500
Location
Icelandic wastelands of Monico, WI
That's because that data should be readily evident in the data sheet .... minus the data on color shift.

He's referring to the NON DATA SHEET info that many of us collect/post. Just like with the xml2 data sheet saying max of 3A but can be driven upwards of 6A with proper cooling etc. Not everyone sees things in the black & white that you do.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
With my industrial experience always focused on reliability and proven performance, I tend to take SemiMan's perspective that if it isn't in the datasheet, then the manufacturer doesn't make any promises. Actually, if the datasheet just lists a "typical" figure, then the manufacturer still isn't making a promise.

Clearly, though, if the OP is buying parts off of ebay, then this isn't an issue of guaranteed performance or quality or lifetime. This is likely a question of how much you can hotrod the LED before it blows up.

It's a bit sad, but CPF used to have a number of people running very interesting experiments on how to improve the thermal transfer of a LED and get more light out of it. Those days seem to be gone, and the OP may have to run a few experiments himself. Take good notes, since killing a few $50 LED modules will be painful!
 

evilc66

Enlightened
Joined
Jul 22, 2008
Messages
276
Chip On Board. They refer to the larger multi-die arrays with a common phosphor layer and flat emitting surface.
 

bhvm

Newly Enlightened
Joined
Jan 24, 2009
Messages
137
He's referring to the NON DATA SHEET info that many of us collect/post. Just like with the xml2 data sheet saying max of 3A but can be driven upwards of 6A with proper cooling etc. Not everyone sees things in the black & white that you do.
Thats what I wanted. Thanks for explaining.

With my industrial experience always focused on reliability and proven performance, I tend to take SemiMan's perspective that if it isn't in the datasheet, then the manufacturer doesn't make any promises. Actually, if the datasheet just lists a "typical" figure, then the manufacturer still isn't making a promise.

Clearly, though, if the OP is buying parts off of ebay, then this isn't an issue of guaranteed performance or quality or lifetime. This is likely a question of how much you can hotrod the LED before it blows up.

It's a bit sad, but CPF used to have a number of people running very interesting experiments on how to improve the thermal transfer of a LED and get more light out of it. Those days seem to be gone, and the OP may have to run a few experiments himself. Take good notes, since killing a few $50 LED modules will be painful!

That made me really nostalgic!
I remember LED gurus like Don Klipstein, Criag Jhonson and "Odd ones' White LED info" website were goldmine for experiments and data. People like don have run EXTENSIVE test on like Likes of Seoul P4 and Luxeon 1~Lux 3. I am missing the pleasure of reading such data on newer COB LEDs. I guess I am going to mail Don straight away.
 

FRITZHID

Flashlight Enthusiast
Joined
Jun 20, 2011
Messages
2,500
Location
Icelandic wastelands of Monico, WI
I think a big reason you haven't found allot of info is simply the cost. Poof a few pricy COBs and you'll get over the idea of testing their limits after a while. It's one thing to poof a few $3-5 LEDs but those $10-50 units add up fast.
More so when output/overdriven input isn't significantly higher then recommended specs.
In my personal exp with COB LEDs, they seem more subject to thermal degradation beyond spec input. I've noticed the phosphor tends to burn very easily causing discoloration and serve color shift.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
The issue with EBag COBs is die matching is very poor and batch to batch conformity is nill. You are literally getting the waste that does not get used elsewhere for the most part.

As well, there operation relies on close matching which is only close over a narrow range. Push the current beyond the data sheet and string matching between strings in the COB gets poor, strings current hog .... And poof.

The cheap COB are already overdriven making things worse.

Posted by really crappy Tapatalk app that is questionable wrt respect of personal data.
 

DIWdiver

Flashlight Enthusiast
Joined
Jan 27, 2010
Messages
2,725
Location
Connecticut, USA
Ditto to SemiMan.

If you are buying LEDs on eBay, you should expect them to be crap. They probably won't meet spec on several key characteristics. They almost certainly won't last like you expect LEDs to last, especially if you drive them near or above ratings.

That said, I might buy LEDs on eBay if:
- I don't expect top performance.
- I plan to underdrive them significantly.
- I don't care if their half-life is 3000 hours instead of 30,000 or 70,000 hours.

If you want good parts, buy from a reputable dealer.
 
Top